Metabolic–inflammatory axis linking diabetes and sarcopenia: cross-population evidence and explainable ai-based risk modeling
Acta Diabetologica · 22.06.2026 · 63 (8) · 1411-1420
DOI: 10.1007/s00592-026-02717-1 ↗Kurz gefasst
Fachartikel in der Zeitschrift Acta Diabetologica, online erschienen am 22.06.2026. Verfasst von Lei He und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Physical Education, Woosuk University, Wanju-gun, 55338, Jeollabuk-do, Republic of Korea. Hauptthemen laut MEDLINE: Sarcopenia, Inflammation, Diabetes Mellitus. Die Arbeit nennt 29 Literaturangaben. Gefördert von: Anhui Provincial Quality Engineering Project, 2025 Research Project on Psychological Health Education in Higher Vocational Colleges, Ministry of Education Vocational Colleges Cultural Quality Education Ste…, Key scientific research project of higher education institutions in Anhui Province.
Themen: Diabetes mellitus
Abstract (Originalsprache)
BACKGROUND: Diabetes and sarcopenia frequently coexist in aging populations and share intertwined metabolic and inflammatory pathways. However, the mechanisms underlying their coexistence and the ability to predict functional decline across populations remain poorly understood. METHODS: Data were obtained from two nationally representative cohorts-NHANES (19,931 participants) and CHARLS (96,628 participants). Diabetes and sarcopenia were defined according to international standards, yielding four groups: control, diabetes only, sarcopenia only, and comorbidity. Metabolic-inflammatory profiles were characterized by CRP, HDL, triglycerides, albumin, and BMI. Cross-sectional and longitudinal analyses were performed to evaluate metabolic-inflammatory-muscle interactions. Logistic, LASSO, and XGBoost models were trained in NHANES to predict functional decline, and their performance and feature importance patterns were evaluated for consistency in CHARLS, with model calibration, decision-curve, and SHAP-based explainability. RESULTS: Individuals with diabetes-sarcopenia comorbidity exhibited the most adverse metabolic-inflammatory pattern, characterized by higher CRP and triglycerides but lower albumin and HDL. Functional performance declined progressively from control to comorbidity groups. The XGBoost model achieved robust discrimination for functional decline (AUC = 0.80 in NHANES; 0.78 in CHARLS) and consistent calibration across subgroups. SHAP interpretation identified albumin, CRP, and HDL as dominant predictors. Mechanistic correlation analysis revealed strengthened negative coupling between glucose, inflammation, and muscle indices in the comorbidity phenotype, supporting an observed association of the metabolic-inflammatory-muscle axis. CONCLUSIONS: Across two independent populations, diabetes-sarcopenia comorbidity was characterized by a convergent metabolic-inflammatory burden that predicted functional decline. The integrated model provided explainable and generalizable risk estimation, highlighting the metabolic-inflammatory-muscle axis as a potentially relevant associated pathway.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Physical Education, Woosuk University, Wanju-gun, 55338, Jeollabuk-do, Republic of Korea.
- School of Physical Education, Chizhou University, No. 199, Mu Road, Guichi District, Chizhou City, 11306, Anhui Province, China.
- Modern Service Academy, Hefei College of Finance & Economics, Hefei, 230601, Anhui, China. 19154011075@163.com.
Förderung
- Anhui Provincial Quality Engineering Project
- 2025 Research Project on Psychological Health Education in Higher Vocational Colleges, Ministry of Education Vocational Colleges Cultural Quality Education Ste…
- Key scientific research project of higher education institutions in Anhui Province
Zitieren
APA
He, L., Wang, S., Cao, Y., Zhang, S., Li, J., Zhao, Z., & Yue, S. (2026). Metabolic–inflammatory axis linking diabetes and sarcopenia: cross-population evidence and explainable ai-based risk modeling. Acta Diabetologica, 63(8), 1411-1420. https://doi.org/10.1007/s00592-026-02717-1
Vancouver
He L, Wang S, Cao Y, Zhang S, Li J, Zhao Z, et al. Metabolic–inflammatory axis linking diabetes and sarcopenia: cross-population evidence and explainable ai-based risk modeling. Acta Diabetologica. 2026;63(8):1411-1420. doi:10.1007/s00592-026-02717-1
BibTeX
@article{He2026,
author = {He, Lei and Wang, Shiping and Cao, Yuang and Zhang, Shihong and Li, JiaJia and Zhao, Zhenhong and Yue, Shun},
title = {Metabolic–inflammatory axis linking diabetes and sarcopenia: cross-population evidence and explainable ai-based risk modeling},
journal = {Acta Diabetologica},
year = {2026},
volume = {63},
number = {8},
pages = {1411--1420},
doi = {10.1007/s00592-026-02717-1}
}
EUVIMED